Fearful Features: Making Risk Explicit

Tell me, Clarice, are the features still screaming?

Published: Tuesday, June 24, 2014 - 11:56

Sometimes work is terrifying. We’re asked to do something that we know, or at least strongly suspect, won’t work. Maybe we’re asked to do several things, some of which are are simple, but one or two make us uncomfortable. We might even place a hard estimate on a task (“It’ll take six hours!”), but deep down something doesn’t feel right. (“Oh well. I’ll just figure that out later.”)

We spend a lot of time guessing how long our work will take. Psychologists have shown this to be a foolish endeavor. They point to the Planning Fallacy or to Hofstadter’s Law, which both show that we are notoriously bad at estimating complex tasks.

But all our work isn’t complex, so we end up with some pretty reliable estimates and some that are wildly off. Those that are wildly off ruin our otherwise fairly predictable projects.

The tendency here would be to try to get rid of those hard-to-estimate tasks.

Good luck with that.

Certainly, we should always be looking at how long it takes for us to complete things, and refine how we project costs and completion times. But variation and difficult tasks run hand-in-hand with knowledge work. Sometimes, we actually have to think about things.

There may be a way out of this.

Work runs along a continuum that looks something like this:

Work is either not scary at all, kind of scary but in a way that you can shoot it, and scary in a way that you’re just unequipped to handle. Or, as we might call them, poodle, crocodile, and zombie.

We want to know, when we are starting some work, which of the tasks we’ve been given that we're comfortable with (and therefore comfortable with the estimate), or which work scares us and to what degree.

I’ve found that teams will give very definite estimates to tasks they are horribly terrified of actually working on. Level of apprehension in the team is an excellent barometer for risk currently being taken on. Risk, in this case, is fairly well-mapped to taking on complex tasks (in the Cynefin sense). And the good news is that there are ways of dealing with complex tasks.

We can then break our work up into simple or complex tasks that can be undertaken by a person or a team with a generally clear view of what they are to do, and how it will be done. So poodle to light-crocodile tasks can be assigned to people with a modicum of comfort that the time estimate can be met.

An actual “Wall of Redacted Fear.”

Crocodile to full-on zombie levels of apprehension, however, indicate that we require a different approach to completing the task.

Estimates are waste: The first thing to come to terms with when a task comes at you is that it does not respect your guesswork. Your estimate is as useless as a handgun against Godzilla. Or a zombie....

One brain bad, four brains good: Complex problems are complex in their composition, implications, and solution sets. When we run into a complex problem that scares individuals on the team, it’s likely that the solution to the problem isn’t right on the tip of someone’s tongue. (And if it is, this person is likely suffering from belief bias). We therefore want to work on complex problems in groups to foster conversation about the solution as it is happening, and employ multiple points of view to reach a meaningful conclusion.

Experiments precede products: These aren’t tasks we just complete and move on to the next. We need to make sure we actually built the right thing. Each complex problem has multiple potential solutions. We need to test the assumptions inherent in our solution by creating an experiment. This could be a mock-up, a prototype, or even a limited release. Whatever we choose, we need to know that “done” for a complex task means proven.

What all this means

We have two types of work:
• Work we think is safe
• Work we think is risky

We’re still going to be wrong when we sort tasks into these two categories, and some risky work will sneak into the safe work. However, if we understand that our estimates are often blown out of the water by risky work that was identifiable up-front, then we can honestly respond to those risks. Right now, we blithely ignore our discomfort and sally forth.

Let’s be honest and get some real work done. Ask which work items worry people, and give that fear the respect it deserves.

First published May 2014 on the Modus Cooperandi blog.

About The Author

Jim Benson

Jim Benson is the creator and co-author (with Tonianne DeMaria) of the best seller Personal Kanban (Modus Cooperandi Press, 2011) winner of the Shingo Research and Publication Award, 2013. His other books include Why Limit WIP (Modus Cooperandi, 2014), Why Plans Fail (Modus Cooperandi, 2014), and Beyond Agile (Modus Cooperandi Press, 2013). He is a winner of the Shingo Award for Excellence in Lean Thinking, and the Brickell Key Award. Benson and DeMaria teach online at Modus Institute and consult regularly, helping clients in all verticals create working systems. Benson regularly keynotes conferences, focusing on making work rewarding and humane.